Thanks to visit codestin.com
Credit goes to github.com

Skip to content

Milestones

List view

  • Support necessary layers for Deeplab v3+ semantic segmentation networks.

    No due date
    7/11 issues closed
  • No due date
    10/14 issues closed
  • No due date
    4/4 issues closed
  • Support standard metrics and analysis that exist in other packages.

    No due date
    1/3 issues closed
  • Find a way to use a Keras-based model description to drive an LBANN training run. This can occur in one of several ways: 1) Directly support the Keras backend API and use LBANN directly. Note that this is not straight-forward because of the parallel nature (MPI) of LBANN. 2) Create a scripted interface that can take Keras code and specify and launch the same model in LBANN.

    No due date
    2/3 issues closed
  • Co-train two independent neural networks with cross model backprop.

    No due date
    0/3 issues closed
  • Add support for working with text networks and other data types that use techniques like word2vec to project the original data into a encoded representation.

    No due date
    2/2 issues closed
  • Support N-way context prediction unsupervised learning technique. The original technique is described in the following two papers. Add support for this as well as new internal evolutions of the technique. Doersch, C., Gupta, A., & Efros, A. A. (2015, May 19). Unsupervised Visual Representation Learning by Context Prediction. arxiv.org. Noroozi, M., & Favaro, P. (2016, March 30). Unsupervised Learning of Visual Representations by Solving Jigsaw Puzzles. arXiv.org.

    No due date
    3/3 issues closed
  • Implement support for unsupervised feature extraction plus fine-tuning.

    No due date
    2/2 issues closed
  • Tune the performance of the existing CPU+GPUs interaction to get good strong scaling as multiple nodes and GPUs are added.

    No due date
    4/4 issues closed